Vocal learning in songbirds, like speech acquisition in humans, requires a period of sensorimotor integration in which vocalizations are evaluated via auditory feedback and progressively refined to achieve an imitation of a vocal model (speech or bird song). Birdsong is controlled by an interconnected system of highly localized brain nuclei, providing an ideal model for studying mechanisms underlying sensorimotor integration for acquiring complex communication signals. As in humans, basal ganglia circuits of songbirds have a critical role in learned vocalizations. In zebra finches, the cortical region LMANcore (see Fig. 1 for acronyms) provides the output of a basal ganglia loop required for vocal learning, while LMANshell provides the output of a parallel basal ganglia loop to both a integrative feed-forward circuit and a feedback pathway via the cortical region, Ad. Recent evidence suggests that LMANcore has a direct motor role in juvenile birds, and that LMANshell/Ad circuitry may serve as the primary site of comparison of vocal-related feedback to the neural memory of the tutor song (the template). Neurons of LMANcore become selectively tuned to the playback of birdsong, especially the bird's own song (BOS) during the sensitive period for vocal learning. It is thought that BOS tuning may encode a neural representation of the current state of the song. It follows that neurons that are selectively tuned to tutor song (TUT) could be used to compare vocal-related feedback to the tutor song. However, only a small proportion of TUT-selective neurons have been found in LMANcore, and selective responses do not occur until late in the sensorimotor phase. Preliminary results show that a considerably greater proportion of individual juvenile LMANshell neurons are exclusively responsive to TUT than juvenile LMANcore neurons. Importantly, those TUT-responsive neurons are more selectively tuned to TUT in LMANshell than LMANcore. Additionally, tuning to BOS emerges in LMANshell at an early stage of sensorimotor integration while this tuning emerges at a later stage in LMANcore. This suggests that there is a representation of both the tutor song and the bird's own song in LMANshell at the beginning of the sensorimotor integration period, as would be expected for a pathway involved in comparison of BOS to TUT. Interestingly, during the period of sensorimotor learning, LMANshell but not LMANcore neurons are selectively inhibited by BOS, which further reflects the disparate functions of these circuits and may be a component of vocal learning. This proposal will test whether LMANshell/Ad circuitry is a primary site of comparison of auditory feedback to the tutor template during sensorimotor integration. Aim 1 will test the selective tuning of neurons in LMANshell versus LMANcore and how differences in tutor-selectivity and inhibition reflect the different functions of these circuits. Aim 2 will test if LMANshell/Ad circuitry has a comparator function by directly perturbing Ad activity during singing.